parsnip

Function jaccard_similiarity_score

Source 
pub fn jaccard_similiarity_score<T>(
    pred: &[T],
    actual: &[T],
) -> Result<f32, LengthError>
where
    T: Eq,
Expand description

The jaccard similarity of a dataset

Returns the jaccard similarity score which for our purposes is effectively categorical accuracy [0, 1]

Supports macro and weighted averages

use parsnip::jaccard_similiarity_score;
let actual = vec![0, 2, 1, 3];
let pred = vec![0, 1, 2, 3];

assert_eq!(jaccard_similiarity_score(&pred, &actual)?, 0.5);